Carbon capture system multi-dimensional interactive design method based on AI technology

Abstract

The invention discloses a multi-dimensional interactive design method for a carbon capture system based on an AI technology, relates to the technical field of computer aided design, and aims to more comprehensively and accurately analyze the performance of the carbon capture system so as to provide a reliable basis for engineering design and ensure the stability and high efficiency of the system in actual operation. The scheme mainly comprises the following five steps: S1, data integration and preprocessing; s2, simulation of an absorption and desorption mechanism; s3, performing fluid dynamic simulation; s4, material and energy balance simulation; and S5, carrying out multi-dimensional interaction and evaluation. The AI technology is combined with a carbon capture system design system, so that the intelligence and high efficiency of the design process are realized. The design efficiency is improved, the design period is shortened, a better design scheme can be obtained, and the performance and reliability of the carbon capture system are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer-aided design. BACKGROUND

[0002] Carbon capture system is a system that captures and processes carbon dioxide through technical means. With the increasing global attention to carbon emission reduction, the accuracy and reliability of the system design of carbon capture technology, as one of the important means to reduce greenhouse gas emissions, are crucial. Traditional carbon capture system design methods often focus on single-dimensional analysis and calculation, such as directly calculating key parameters such as flue gas volume and circulation volume through material balance software, and then carrying out process selection and design, and lack of interactive feedback mechanism, which is difficult to fully and accurately reflect the performance and problems of the system in actual operation, resulting in the designed system may have defects such as unstable operation and high energy consumption.

[0003] The Chinese patent with patent publication number CN117875167B is a method, device and equipment for optimizing process parameters of a flue gas CO2 capture device. The method includes: based on the real-time measurement data of the CO2 capture device and the pre-trained gated recurrent neural network model, predicting the key operation parameter data of the offline laboratory analysis in the amine liquid system of the CO2 capture device; using the predicted key operation parameter data of the offline laboratory analysis and the real-time measurement data as the basis for automatic control adjustment and operator adjustment of the CO2 capture device to optimize and control the process of the CO2 capture device in real time. This scheme has certain value in improving the running efficiency of the device, but it still has obvious limitations. This method mainly relies on a single type of AI model and focuses on local parameter optimization of the amine liquid system, lacks multi-dimensional modeling and collaborative optimization of the overall performance of the carbon capture system, and fails to comprehensively consider key factors such as absorption and desorption mechanism, fluid dynamics behavior and material energy balance, so there are still certain shortcomings in system complexity, design comprehensiveness and engineering adaptability.

[0004] Any discussion of background art throughout the specification should in no way be considered as an admission that the background art was prior art to the present application; any discussion of the prior art throughout the specification is intended merely to illustrate that the present application is not limited to a specific aspect of the prior art. SUMMARY

[0005] The present application aims to provide a multi-dimensional interactive design method for carbon capture systems based on AI technology, which can more comprehensively and accurately analyze the performance of carbon capture systems, thereby providing reliable basis for engineering design and ensuring the stability and efficiency of the system in actual operation.

[0006] The multi-dimensional interactive design method for carbon capture systems based on AI technology in this scheme includes the following steps: S1: Data integration and preprocessing; S11: Confirming design targets and requirements for the carbon capture system under construction; S12: Collecting project base parameters and design indicators; S13: Cleaning and standardizing the project base parameters and design indicator data; S2: Absorption and desorption mechanism simulation; S21: Establishing an absorption and desorption mechanism model in a simulation environment, inputting the project base parameters and design indicators obtained in S1 into the absorption and desorption mechanism model, and obtaining absorption and desorption simulation data and a first version of design parameters, wherein the desorption simulation data includes state parameters and required resource amounts at each point in the carbon capture system; S22: Establishing an absorption and desorption mechanism AI model to predict and optimize the parameters in the absorption and desorption mechanism model; S3: Fluid dynamics simulation; S31: Establishing a fluid dynamics model, inputting the absorption and desorption simulation data, and obtaining fluid dynamics simulation data; S32: Inputting the fluid dynamics simulation data into a fluid dynamics AI model to obtain optimized operating parameters, inputting the first version of design parameters and the optimized operating parameters into the absorption and desorption mechanism model and the fluid dynamics model to form a second version of optimized design parameters; S4: Material and energy balance simulation; S41: Establishing a material and energy balance model, inputting the second version of design parameters, and obtaining material and energy simulation data; S42: Inputting the material and energy simulation data into a material and energy balance AI model for optimization; S5: Multi-dimensional interaction and evaluation; S51: Establishing a multi-dimensional interaction AI model, inputting design indicators, absorption and desorption simulation data, fluid dynamics simulation data, material and energy simulation data, and the second version of design parameters, and outputting a third version of design parameters and performance evaluation; S52: If the performance evaluation does not meet the design targets and requirements, the AI model is allowed to optimize and output a new version of design parameters and performance evaluation.

[0007] The design indicators include flue gas parameters (temperature, pressure, and content of each component such as CO2 and O2 in the flue gas), absorbent performance parameters (concentration, absorption and desorption performance, and thermal stability), and device operating conditions such as temperature, pressure, flow rate, and liquid level of devices such as absorption towers, desorption towers, heat exchangers, and reboilers.

[0008] The state parameters include flue gas flow rate, absorbent flow rate, temperature, and pressure; and the required resource amounts include power and steam amounts.

[0009] The preferred operating parameters include the velocity distribution, concentration distribution and heat and mass transfer process of the absorbent, flue gas, etc. in the tower.

[0010] The design parameters include the tower diameter, packing layer, absorbent circulation amount, inlet inclination angle, etc. of the absorption tower, and also include the operating parameters.

[0011] The material and energy simulation data include the CO2 absorption amount in the absorption tower, the solvent circulation amount in the system, the steam consumption, etc.

[0012] The effective effects of the present scheme are as follows: (1) AI-driven multi-software collaborative optimization: The absorption and desorption mechanism simulation can be completed using Aspen software, the fluid dynamics simulation can be realized through Fluent software, and the material and energy balance simulation can be completed using related software. Through AI technology, multiple professional software such as Aspen, Fluent and material and energy balance software are organically coupled together, realizing the full-process intelligent collaborative optimization from mechanism model establishment to fluid dynamics simulation and then to material and energy balance control. The performance indicators of the system can be comprehensively evaluated and analyzed, fully considering the mutual influence between various factors, and providing a scientific basis for further optimization of the design scheme.

[0013] (2) Intelligent parameter optimization and feedback mechanism: Using the machine learning and deep learning algorithms of AI, the simulation results of each software are analyzed and optimized, the best design parameters are automatically determined, and the design scheme is continuously adjusted and improved through the feedback mechanism.

[0014] (3) Full-process intelligent regulation and control and problem early warning: Combined with the prediction and optimization functions of AI, the full process of the carbon capture system is intelligently regulated and controlled, problems that may occur in the system are discovered and corrected in a timely manner, and the stable operation of the system is ensured.

[0015] The present application realizes the intelligentization and high efficiency of the design process by combining AI technology with the carbon capture system design system. Not only the design efficiency is improved and the design cycle is shortened, but also a better design scheme can be obtained, and the performance and reliability of the carbon capture system are improved.

[0016] Further, it further includes S6: generation and optimization of engineering design documents: S61: establishing a document template, which includes the structure, format and content requirements of the document; S62: inputting the absorption and desorption simulation data, fluid dynamics simulation data, material and energy simulation data, design parameters and document template into the AI model, and outputting the engineering design document; S63: inputting the existing normative files into the AI model, outputting the modification suggestions of the engineering design document, or directly outputting the modified engineering design document.

[0017] The scheme can greatly improve the production efficiency of engineering design documents, and also ensure the quality and operability of engineering design documents. The intelligently generated engineering design documents also provide accurate guidance for engineering implementation, with significant economic benefits.

[0018] Further, the design indicators collected in S12 include: Flue gas parameters, temperature, pressure and / or component content; Absorbent performance parameters: concentration, absorption and desorption performance and / or thermal stability; Device operating conditions, temperature, pressure, flow rate and / or liquid level. The device includes an absorption tower, a desorption tower, a heat exchanger, etc.

[0019] Further, the AI model established in S22 is a machine learning model, and the built-in methods include random forest regression algorithm, neural network and support vector machine, which are used to predict and optimize the key parameters in the absorption and desorption mechanism model.

[0020] Further, the fluid dynamics model established in S31 adopts RNG turbulence model and SIMPLE algorithm, and combines DPM discrete phase model to model the liquid droplet spraying process.

[0021] Further, the AI model in S32 uses convolutional neural network or recurrent neural network to analyze the fluid dynamics simulation data to identify the key features of velocity distribution and concentration distribution and output fluid flow parameters.

[0022] Further, the material and energy balance AI model in S42 is a data-driven model based on machine learning algorithm, which has prediction and optimization functions, and is used to control the material input, output and energy conversion of the system.

[0023] Further, the multi-dimensional interactive AI model established in S51 has three layers of data layer, model layer and application layer, wherein the model layer is used to manage the calculation results of Aspen process simulation model, Fluent fluid dynamics model and material and energy balance model.

[0024] Further, when the AI model in S52 generates new design parameters based on performance evaluation results, a reinforcement learning algorithm is used for multiple rounds of iterative optimization until the design target and system performance requirements are met. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 The figure is a schematic diagram of the system architecture of the embodiment of the present application.

[0026] Figure 2 The figure is a design flowchart of the embodiment of the present application.

[0027] Figure 3Schematic diagram of the architecture of the multi-dimensional interactive platform in an embodiment of the present invention.

[0028] Figure 4 Schematic diagram of AI model training and optimization in an embodiment of the present invention.

[0029] Figure 5 A flow chart for generating engineering design documents in an embodiment of the present invention. DETAILED DESCRIPTION

[0030] The present invention will be further described in detail below through specific embodiments: The overall overview of the embodiment is basically as shown in the attached Figure 1 、 2 As shown: This example combines various professional tools, such as Aspen, Fluent, and material energy balance software, with AI technology to achieve intelligent and efficient carbon capture system design. The specific technical solution is as follows.

[0031] (1) Step 1: Data integration and preprocessing For a certain carbon capture system, the design objectives and requirements are clearly defined, various data in the carbon capture project are collected, including flue gas parameters, absorbent performance parameters, capture capacity and equipment operating conditions, etc., and data cleaning and standardization are carried out.

[0032] 1) Data collection: covers flue gas parameters (temperature, pressure, content of CO2, O2 and other components in flue gas), absorbent performance parameters (concentration, absorption and desorption performance, thermal stability, etc.), and equipment operating conditions such as temperature, pressure, flow rate, liquid level of absorption towers, desorption towers, heat exchangers, and reboilers.

[0033] 2) Data cleaning and standardization: Remove outliers, missing values, and duplicate data to ensure data accuracy and completeness; standardize data from different sources and units to unify the data dimensions and units.

[0034] (2) Step 2: Establishment and optimization of AI-driven absorption and desorption mechanism model 1) Based on the absorption and desorption performance of the high-efficiency absorbent, the absorption and desorption mechanism model was established using Aspen software. The project-related input parameters obtained in step 1 were input to calculate the state parameters of each point in the system and the amount of electricity and steam required by the system, as follows: Establishing an Aspen absorption and desorption model: In Aspen, the process flow is established, operating conditions are set, and the absorbent's absorption and desorption properties are input to create a carbon capture absorption and desorption mechanism model. After the process simulation model is established, the project's capture capacity and flue gas composition parameters are input to calculate state parameters such as flue gas flow rate, absorbent flow rate, temperature, and pressure at each point in the system, as well as the system's required power and steam.

[0035] 2) Use the collected data to train the machine learning algorithm, establish the AI model, and optimize the parameters of the Aspen absorption and desorption model, as follows: a) AI model establishment: First, collect other project case data and input it into the established model to obtain multiple historical data. Second, simulate multiple projects and collect historical calculation data of the Aspen model. Use the input-output relationship in the historical data to establish an AI model, train the model to predict system performance indicators such as capture efficiency and energy consumption using methods such as random forest regression, neural networks, and support vector machines, and provide a basis for optimizing operation. This model can automatically learn the characteristics and rules in the data of each project and optimize the parameters of the Aspen model.

[0036] b) Model optimization and feedback: Feedback the optimization results of the AI model to the Aspen process simulation model to adjust key parameters such as the reaction rate of the absorbent, the equilibrium constant, and the operating pressure of the regeneration tower, and improve the accuracy of the model.

[0037] (3) Step three: Fluid dynamics simulation and AI analysis 1) Use the RNG model (turbulence model), SIMPLE algorithm (velocity pressure coupling), and discrete phase model (DPM, liquid droplet spraying) to describe the flue gas in the calculation region as a continuous medium in the Euler coordinate system, solve the equations for the continuous phase, and solve the particle trajectory of the discrete phase by integrating the particle force differential equation in the Lagrangian coordinate system. Establish a full-flow two-phase flow Fluent fluid dynamics model for the absorption and desorption system, input the state parameters obtained from the Aspen process simulation calculation into the software, solve the fluid dynamics equations, and obtain the velocity distribution, concentration distribution, and heat and mass transfer process of the absorbent and flue gas in the tower.

[0038] 2) Feedback the data obtained from each Fluent simulation to the AI model and use deep learning algorithms to analyze the data to determine the optimal fluid flow parameters. Further, feedback to the Aspen process simulation model for further process simulation calculation, and modify the key structure and operation parameters of the absorption tower, such as tower diameter, packing layer, absorbent circulation amount, and inlet inclination angle, as follows: a) Fluent simulation: Input the flow rate, temperature, and other state parameters of the flue gas and absorbent obtained from the Aspen process simulation calculation, as well as the tower diameter of the absorption tower and regeneration tower, and packing parameters into the two-phase flow dynamics model to solve the fluid dynamics equations and obtain the velocity distribution, concentration distribution, and heat and mass transfer process of the absorbent and flue gas in the tower.

[0039] b) Continuously adjust the tower diameter, inlet angle, liquid-gas ratio and other parameters, and automatically transmit the data obtained from each simulation to the AI model through the multi-dimensional interactive platform. Use deep learning algorithms such as convolutional neural networks (CNN) or recurrent neural networks (RNN) to analyze the velocity distribution, concentration distribution and other data. The AI model can automatically identify key features and patterns in the data and determine the best fluid flow parameters, such as the optimal flue gas flow rate, inlet angle and optimal liquid-gas ratio.

[0040] c) Tower structure optimization: Feedback the optimal parameters obtained from AI analysis to the Aspen process simulation model and Fluent dynamics model, and modify the tower diameter, filler layer, absorbent circulation amount, inlet angle and other key structure and operation parameters of the absorption tower to improve mass transfer efficiency and reduce system energy consumption.

[0041] (4) Step four: Intelligent control of material and energy balance in the whole process 1) Based on the existing key parameters of absorption and desorption, use Excel macro tools to create material and energy balance analysis software to calculate and analyze the material flow and energy change in the production process. Input the optimized state parameters (such as flue gas volume, tower diameter, etc.) obtained from steps two and three into the material and energy balance software to calculate the input, output and storage of materials in the system throughout the process, ensuring that the system maintains a stable state under different operating conditions. Calculate the CO2 absorption amount in the absorption tower, the circulation amount of the solvent in the system, the steam consumption and other key data, and further compare and correct the circulation amount, steam consumption and other key data obtained from steps two and three.

[0042] 2) Embed AI prediction and optimization functions to intelligently control the material input, output and energy conversion of the system. The embedded AI model can predict the required absorbent amount and steam volume according to the real-time flue gas flow and composition changes that may occur during actual operation, and provide suggestions for adjusting operating parameters. Finally, output the system operating parameters and operation suggestion documents under different conditions to ensure that the system operates in a stable, efficient and energy-saving state.

[0043] (5) Multi-dimensional interactive mechanism and comprehensive evaluation of system performance Establish a multi-dimensional interactive platform based on AI technology to integrate and interactively analyze the calculation results of the Aspen process simulation model, Fluent fluid dynamics model and material and energy balance software. Use the pattern recognition and data analysis capabilities of AI to comprehensively evaluate and analyze the performance indicators of the system, and further optimize the design scheme based on the evaluation results. The overall architecture is shown in Figure 3 , 4 .

[0044] The main structure is as follows: 1) Platform Architecture: including data layer, model layer and application layer. The data layer is responsible for collecting project data; the model layer is responsible for managing the calculation results of the three models of Aspen process simulation model, Fluent fluid dynamics model and material energy balance software, and analyzing, predicting and optimizing the data; the application layer provides multi-dimensional interaction of the three models, and makes system performance evaluation and output design documents.

[0045] 2) Platform Interface: including data interface, model interface and application interface. The data interface is used to provide basic data for the data layer; the model interface provides an interface for integrating Aspen process simulation model, Fluent fluid dynamics model and material energy balance software into the AI platform, facilitating the training, deployment and calling of the model; the application interface supports the development and integration of various application functions, such as system performance evaluation, engineering design document generation, etc.

[0046] The implementation of the functions is as follows: 1) Multi-software collaborative simulation and analysis: comprehensive analysis of the simulation results of Aspen, Fluent and material energy balance software, etc., correlation and fusion of data between different software through AI model, realization of multi-physical field, multi-scale collaborative simulation. More specifically: the AI model connects Aspen, Fluent and material energy balance software through standardized interface, obtains the output data of each model, and after format unification and unit conversion, carries out cross-model data fusion, and based on the fused data, carries out multi-physical field coupling calculation, realizes multi-scale collaborative simulation of heat transfer, mass transfer, fluid flow, etc.

[0047] 2) System performance evaluation and optimization: based on AI model and multi-dimensional data analysis technology, the performance indicators of carbon capture system are comprehensively evaluated, such as capture efficiency, energy consumption, equipment utilization, etc. According to the evaluation results, the platform can automatically propose optimization suggestions and schemes to guide users to improve and optimize the system.

[0048] (6) Intelligent generation and optimization of engineering design documents, the process is as shown in Figure 5

[0049] 1) According to the optimized design parameters and simulation results, use AI technology to automatically generate detailed engineering design documents, including system layout diagram, equipment selection, operation manual, etc., as follows: a) Document template establishment: according to the requirements of carbon capture project and industry standards, establish the template of engineering design documents, including system layout diagram, equipment selection table, operation manual, etc.; the template defines the structure, format and content requirements of the document, providing the basis for intelligent generation of documents.

[0050] ​b) Intelligent document generation: automatically transfer the optimized design parameters and simulation results to the document generation system, use natural language processing (NLP) and automatic drawing algorithm to automatically generate detailed engineering design documents; the system can automatically fill in relevant data and information according to the requirements of the template, and generate accurate and complete document content.

[0051] 2) Use AI rule engine and machine learning algorithm to intelligently review and optimize the generated engineering design documents. The AI system can check the data consistency, logical correctness and compliance with industry standards in the document, and automatically correct the problems found. At the same time, according to the optimization suggestions of the system, the design scheme in the document is further improved to ensure the quality and operability of the engineering design document.

[0052] Specific examples: In a certain carbon capture project, the design goal is to capture 10t / h and the capture efficiency is 90%, and the system energy consumption requirement reaches the industry leading level. According to these requirements, the basic configuration of the carbon capture system is determined, including 1 set of washing tower, 1 set of absorption tower, 1 set of desorption tower, 3 sets of heat exchanger, 1 set of reboiler, 5 sets of circulating pump, 1 set of compression drying system, 1 set of refrigeration liquefaction system and other equipment and its operating conditions.

[0053] (1) Step one: data integration and preprocessing Collect the data of this project, including the capture amount of the project 10t / h, the capture efficiency 90%, the flue gas parameters (temperature 58℃, pressure 0.8kpa, CO2 content 12%, O2 content 5%, H2O content 14%, N2 content 69% and other data), steam parameters (temperature 135℃, pressure 0.4MPa (gauge pressure)) and other data. The collected data is cleaned and standardized, such as the conversion of flue gas standard flow and actual flow, the reasonableness check of flue gas components, etc., to remove outliers, missing values and repeated data, etc., to ensure the accuracy and integrity of the data.

[0054] (2) Step two: AI driven absorption and desorption mechanism model establishment and optimization Based on the absorption and desorption performance of high-efficiency absorbent, an absorption and desorption mechanism model is established using Aspen software, which automatically imports the relevant parameters of the project in step one, obtains the gas-liquid phase equilibrium, mass balance and energy balance parameters, and calculates the flue gas flow at each point in the system, such as the flue gas flow into the washing tower 47138.05 Nm 3 / h, the absorbent circulation amount 239m 3 / h, the flue gas entering the absorption tower temperature 40℃, the lean liquid entering the absorption tower temperature 40℃, the absorption tower pressure drop 2.5kpa, the regeneration tower bottom temperature 110℃, the tower top pressure 18kpa, etc. The system requires 1959 degrees / h (including compression liquefaction) of power and 10.91t / h of steam.

[0055] In addition, historical data (other project cases) and simulation results of the Aspen model can be used to train the neural network algorithm. Through continuous iterative training, the AI model can accurately predict and optimize key parameters in the Aspen model, such as the actual absorption rate of the absorbent in the tower 0.196 mol CO2 / (L·min), the desorption rate 0.095 mol CO2 / (L·min), etc. The optimization results of the AI model are fed back to the Aspen model to adjust the model parameters and improve the accuracy and prediction ability of the model.

[0056] (3) Step three: fluid dynamics simulation and AI analysis Using the flue gas flow, absorbent flow, temperature, pressure, and other state parameters obtained in the above steps, the fluid dynamics equations are solved by Fluent software to obtain the velocity distribution, concentration distribution, and heat and mass transfer process of the absorbent and flue gas in the tower. By simulating the fluid behavior at different flow rates and inputting the Fluent simulation data into the AI model, the velocity distribution, concentration distribution, and other data are analyzed using convolutional neural networks (CNN) to determine the optimal flue gas flow rate of 1.52 m / s, the inlet inclination angle of 15°, and the optimal liquid-gas ratio of 4.985 L / m³.

[0057] The optimal parameters obtained by AI analysis are fed back to the Fluent software to modify the tower diameter, filler layer, and absorbent circulation amount of the absorption tower, and determine the tower diameter of the absorption tower as 3.5 m, the filler layer as 3 layers with a total height of 15 m, the tower diameter of the regeneration tower as 1.8 m, the filler layer as 2 layers with a total height of 12 m, and set the liquid collector, gas-liquid distributor, and other devices.

[0058] (4) Step four: intelligent control of material and energy balance in the whole process The flue gas flow, absorbent circulation amount, temperature, pressure, and other state parameters obtained above are input into the material and energy balance software. Through full-process calculation of the input, output, and storage of materials in the system, the system is ensured to maintain a stable state under different operating conditions. The calculation results show that the absorption amount of CO2 in the absorption tower is 10 tons per hour, the desorption amount of CO2 in the desorption tower is 10 tons per hour, and the circulation amount of the solvent in the system is 235 m³ / h. 3 / h (corrected), and promptly identified and corrected potential issues in the system, such as solvent loss and energy imbalance. The required hourly absorbent replenishment was determined to be 3.25 kg, and the required desalinated water replenishment was 32 kg. By feeding these results back into the simulation model for verification and optimization, a multi-dimensional interactive mechanism was established, enabling a comprehensive analysis of system performance. Ultimately, further optimization of system energy and material consumption was achieved, resulting in calculated system power requirements of 1878 kWh / h (including compression and liquefaction), steam production of 10.47 t / h, and absorbent loss of 3.25 kg / h.

[0059] In addition, the prediction and optimization functions of AI can be combined to predict the absorption dose and steam volume required by the system based on the real-time flue gas flow and composition changes that may occur during actual operation, and propose operating parameter adjustment suggestions. Finally, the system operating parameters and operation suggestion documents under different states are output to ensure that the system operates in a stable, efficient and energy-saving state. The predicted load changes are referenced below.

[0060]

[0061] (5) Comprehensive evaluation of multi-dimensional interaction mechanism and system performance A multi-dimensional interactive platform based on AI was established to integrate and interactively analyze the calculation results of Aspen, Fluent, and material energy balance software, realizing real-time sharing and interaction of data in steps (2), (3), and (4). Utilizing AI's pattern recognition and data analysis capabilities, various performance indicators of the system were comprehensively evaluated and optimized, ultimately achieving a carbon capture rate of 93.2%, a regenerative heat consumption of 2.35 GJ / t CO2, an absorbent loss of 0.325 kg / t CO2, and a full liquefaction power consumption of 187.8 kWh / tCO2. By comparing with historical data and industry standards, the technical indicators of this design have reached the industry-leading level.

[0062] (6) Intelligent generation and optimization of engineering design documents Based on the optimized design parameters and simulation results, detailed engineering design documentation is generated, including system layout diagrams, equipment selection, energy consumption analysis, and operating manuals, providing precise guidance for project implementation. The overall efficiency of carbon capture system designs using this method has increased by over 200% compared to manual design, and the accuracy of design parameters has increased by over 50%.

[0063] The above is only the embodiment of the present application, and the technical means not mentioned can use the prior art for the skilled in the art. Without departing from the scheme of the present application, several modifications and improvements can also be made, which should also be considered as the protection scope of the present application, and these will not affect the effect and practicality of the present application. The protection scope claimed by the present application should be subject to the content of its claims, and the specific implementation mode and the like recorded in the specification can be used to explain the content of the claims.

Claims

1. A multi-dimensional interactive design method for carbon capture systems based on AI technology, characterized by The following steps are involved: S1: Data integration and preprocessing; S11: Confirm the design objectives and requirements for the carbon capture system under construction; S12: Collect basic project parameters and design indicators; S13: Clean and standardize the basic parameters and design index data of the project; S2: simulation of absorption and desorption mechanism; S21: Establishing an absorption and desorption mechanism model in a simulation environment, inputting the basic project parameters and design indicators obtained in S1 into the absorption and desorption mechanism model, and obtaining absorption and desorption simulation data and a first version of design parameters. The desorption simulation data includes state parameters and required resource quantities at each point in the carbon capture system; S22: Establishing an AI model of absorption and desorption mechanism, and predicting and optimizing the parameters in the absorption and desorption mechanism model; S3: fluid dynamics simulation; S31: establishing a fluid dynamics model, inputting absorption and desorption simulation data, and obtaining fluid dynamics simulation data; S32: Inputting the fluid dynamics simulation data into the fluid dynamics AI model to obtain the optimal operating parameters, and inputting the first version of the design parameters and the optimal operating parameters into the absorption and desorption mechanism model and the fluid dynamics model to form the optimized second version of the design parameters; S4: Material and energy balance simulation; S41: Establish a material energy balance model and input the second version design parameters to obtain material energy simulation data; S42: Inputting material energy simulation data into the material and energy balance AI model for optimization; S5: Multidimensional Interaction and Assessment; S51: Establish a multi-dimensional interactive AI model, input design indicators, absorption and desorption simulation data, fluid dynamics simulation data, material energy simulation data and the second version of design parameters, and output the third version of design parameters and performance evaluation; S52: If the performance evaluation does not meet the design goals and requirements, let the AI ​​model optimize and output a new version of the design parameters and performance evaluation.

2. The multi-dimensional interactive design method for carbon capture system based on AI technology according to claim 1 is characterized in that Also includes S6: Generation and optimization of engineering design documents: S61: Create a document template, which includes the document's structure, format, and content requirements; S62: Input absorption and desorption simulation data, fluid dynamics simulation data, material energy simulation data, design parameters and document templates into the AI ​​model, and output engineering design documents; S63: Input existing regulatory documents into the AI ​​model and output modification suggestions for the engineering design document, or directly output the modified engineering design document.

3. The AI-based multi-dimensional interactive design method for carbon capture systems according to claim 2, characterized in that: The basic project parameters and design indicators collected in S12 include: Flue gas parameters, temperature, pressure and / or component content; Absorbent performance parameters: concentration, absorption and desorption performance and / or thermal stability; Equipment operating conditions, temperature, pressure, flow and / or level.

4. The AI-based multi-dimensional interactive design method for carbon capture systems according to claim 3, characterized in that: The AI ​​model established in S22 is a machine learning model with built-in methods including random forest regression algorithm, neural network and support vector machine, which is used to predict and optimize key parameters in the absorption and desorption mechanism model.

5. The AI-based multi-dimensional interactive design method for a carbon capture system according to claim 4, characterized in that: The fluid dynamics model established in S31 uses the RNG turbulence model and SIMPLE algorithm, combined with the DPM discrete phase model to model the droplet spraying process.

6. The AI-based multi-dimensional interactive design method for a carbon capture system according to claim 5, characterized in that: The AI ​​model in S32 uses convolutional neural networks or recurrent neural networks to analyze fluid dynamics simulation data to identify key features of velocity distribution and concentration distribution and output fluid flow parameters.

7. The AI-based multi-dimensional interactive design method for a carbon capture system according to claim 6, characterized in that: The material and energy balance AI model in S42 is a data-driven model built based on machine learning algorithms. It has prediction and optimization capabilities and is used to regulate the system's material input, output, and energy conversion.

8. The AI-based multi-dimensional interactive design method for a carbon capture system according to claim 7, characterized in that: The multi-dimensional interactive AI model established in S51 integrates a three-layer architecture consisting of data layer, model layer and application layer. The model layer is used to manage the calculation results of the Aspen process simulation model, Fluent fluid dynamics model and material energy balance model.

9. The AI-based multi-dimensional interactive design method for a carbon capture system according to claim 8, characterized in that: When the AI ​​model in S52 generates new design parameters based on performance evaluation results, it uses a reinforcement learning algorithm to perform multiple rounds of iterative optimization until the design goals and system performance requirements are met.

Images